It’s been awhile since my last article on Medium. Life has been all over the moon for me the past few months, therefore I forgot one of my most important duty here.
This time, I try something different. I wanna compare how good ChatGPT 3.5 and Barb is, when it comes to writing a technical Medium article.
I started with drafting out the summary of what I wanna write about, which is just a simple bullet points.
Details on all Dart keywords and how to use them - import: include the public code from another file into your current file - import show import public code from a file, but show only some specific classes or functions - import hide import public code from a file but exclude some classes or function - export: share the code from the targeted file to any file that import this file - export show: export specific classes or function to any file that import this file - export hide: export all code from this file except some specific classes or functions - part: include source code from another file (public or private) as part of this file - part of: declare this file as a part of another file. ‘part’ and ‘part of’ are a pair , and go together - mixin - extension - const - extends - implement - async - await - void - final - abstract - Typedef - @overrides - @visibleForTesting - dynamic - var - ? - ?? - !
As you can see from my simple note above, there are a list of dart keywords, with little to no explanation of them. I send this simple note to both ChatGPT and Bard, with the same request
Complete this article for me, with code sample
The result from ChatGPT is a 10 page in details of each keyword, with code sample how to use them, and I have to say that the information for this post is all correct, with detail example in Medium ready format.
In the other hand, Bard from Google give me a short 3 pages answer with some (but not all) code sample for the same request, and with very brief explanation for each keyword.
However, Bard gives a much better code UI display, with keyword colouring and all. And what’s best, Bard gives me 3 drafts to choose from. Really impressive.
Conclusion
In this test, I think both ChatGPT and Bard have done well. They both give accurate details and adequate code sample. However ChatGPT gives more details on each keyword, and more code sample, 1 for each keyword, whereas Bard is kinda abstract.
For the knowledge, I’ll paste the full answer from both ChatGPT, and Bard’s answer can be found in this link. It’s always good to learn something, from AI or human alike.
Comments